In patients with heart failure (HF), accurate differentiation among ischemic and various non-ischemic cardiomyopathies (NICMs) is essential for appropriate guidance of therapies. Late gadolinium enhancement (LGE) MRI is an ideal non-invasive imaging method for depicting myocardial scarring and fibrosis, but is not well suitable for HF patients due to the requirement of many breath-holds and limited specificity to distinguish various forms of NICM. While LGE MRI is based on static analysis of the concentration of contrast agent at a single post-injection time, contrast uptake and washout is actually a temporally dynamic process, the pattern of which varies depending on specific pathological conditions. If accurately captured, the contrast kinetics may improve the tissue characterization particularly for differentiating NICMs that are associated with different etiologies. The ultimate goal of this project is to develop so-called early-to-late gadolinium enhancement (ELGE) MRI methods which capture contrast uptake and wash-out over time and to utilize the resultant kinetics information for differentiating various NICMs. Towards this goal, we will first develop respiratory motion corrected 3D LGE imaging methods which enable whole heart coverage during free-breathing. This will be achieved by integrating novel MRI components such as 3D stack-of-spirals acquisition, outer volume suppression magnetization preparation and 1D projection-based motion estimator. Then, ELGE measurement will be done in NICM patients by repeating the free-breathing scan between 1min through 40 min post injection. After parameterization of ELGE time curves, classification rules will be established to distinguish myocarditis, dilated cardiomyopathy and sarcoidosis that justify a larger clinical study.